Stop Throwing Away Discriminators! Re-using Adversaries for Test-Time Training

  title={Stop Throwing Away Discriminators! Re-using Adversaries for Test-Time Training},
  author={Gabriele Valvano and Andrea Leo and Sotirios A. Tsaftaris},
Thanks to their ability to learn data distributions without requiring paired data, Generative Adversarial Networks (GANs) have become an integral part of many computer vision methods, including those developed for medical image segmentation. These methods jointly train a segmentor and an adversarial mask discriminator, which provides a data-driven shape prior. At inference, the discriminator is discarded, and only the segmentor is used to predict label maps on test images. But should we discard… Expand

Figures from this paper


Learning from Simulated and Unsupervised Images through Adversarial Training
This work develops a method for S+U learning that uses an adversarial network similar to Generative Adversarial Networks (GANs), but with synthetic images as inputs instead of random vectors, and makes several key modifications to the standard GAN algorithm to preserve annotations, avoid artifacts, and stabilize training. Expand
Is Discriminator a Good Feature Extractor?
The performance of the discriminator extractor may be inferior to the classifier based extractor when the source classification task is similar to the target task, which is the common case, but the ability to avoid noise prevents the discriminators from being replaced by the classifiers. Expand
Memorization Precedes Generation: Learning Unsupervised GANs with Memory Networks
It is argued that these two infamous problems of unsupervised GAN training can be largely alleviated by a learnable memory network to which both generators and discriminators can access, and a novel end-to-end GAN model named memoryGAN is proposed, which involves a memory network that is un supervisedly trainable and integrable to many existing GAN models. Expand
Generative Adversarial Nets
We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and aExpand
Adversarially Learned Anomaly Detection
This work proposes an anomaly detection method, Adversarially Learned Anomaly Detection (ALAD), that derives adversarially learned features for the anomaly detection task that achieves state-of-the-art performance on a range of image and tabular datasets while being several hundred-fold faster at test time than the only published GAN-based method. Expand
Generative Adversarial Network in Medical Imaging: A Review
A review of recent advances in medical imaging using the adversarial training scheme with the hope of benefiting researchers interested in this technique. Expand
A Large-Scale Study on Regularization and Normalization in GANs
This work takes a sober view of the current state of GANs from a practical perspective, discusses and evaluates common pitfalls and reproducibility issues, and open-source the code on Github and provide pre-trained models on TensorFlow Hub. Expand
On the Effectiveness of Least Squares Generative Adversarial Networks
The Least Squares Generative Adversarial Networks (LSGANs) are proposed which adopt the least squares loss for both the discriminator and the generator, and LSGANs are able to generate higher quality images than regular GANs. Expand
Theoretical Insights into Memorization in GANs
While significant research has gone into developing Generative Adversarial Network (GAN) architectures and training procedures that produce high-quality images, very little is known about how andExpand
Autoencoder based self-supervised test-time adaptation for medical image analysis
This paper proposes a model that can adapt based on a single test subject during inference, which can achieve significant performance improvement and was validated on both retinal optical coherence tomography (OCT) image segmentation and magnetic resonance imaging (MRI) T1-weighted to T2- Weighted image synthesis. Expand